Integration Adapting existing technologies to new standards
From "Special Report",  Access to Wang, March 1992
[ Prior Article ]     [ Return to the Catalog of articles ]     [ Next Article ]

Whatever the final destination for your shop, integration with the VS should be considered a requirement for any new systems. Many of us have grown used to the sheltered environment offered the VS and so we anticipate a rough time getting used to a new syntax, new terminals, maybe even a new language or new development approach. Critical business functions should continue to be carried by the existing system until it is clear that whatever change you opt for will be successful.

Strange times are indeed upon us. A struggling Wang Laboratories has turned to marketing IBM equipment for its survival, while other computer equipment vendors - also grappling for a place in a weak market - have invented sometimes improbable schemes to attract disciples to their architectures. Many businesses that have reached the practical limits of the VS architecture and that held out for promised improvements from Wang, now find that the VS itself may be entering the "sunset" phase in its life.

Like so much static on your AM dial, you are unable to evaluate clearly what you need to do and where you need to go. The choice is not a clear one, for through it all, your VS continues to hum away and quietly support your business.

My purpose here is not to debate the wisdom of changes to your system environment - these changes are now inevitable. Instead, I offer one viewpoint on the manner in which new systems should be introduced: through integration and eventual migration.

Parallel Universes

There are reasons for the VS's success and reasons to keep the system. It has been quietly supporting your business needs without requiring a technical staff. Then there are these considerations:

• Over time, the average VS site has responded to the relatively easy way data is shared on the VS. The user base is now immersed in ad hoc queries and other user-controlled uses. Their interest has outstripped the supply of software, and you're spending money developing new systems.

• Programmers consider the VS to be a friendly, integrated environment and are comfortable programming in it.

• There's just too much residual value (and ego) in the current system to toss it away.

• It's so darn dependable.

• But the rose garden does contain thorns.

• While in a holding pattern deciding what to do about the system, the average site has continued to add new users while trying to satisfy increasing user demands on the system. System performance has degraded noticeably.

External evidence of Wang's corporate reversals has been news, so management may be increasingly nervous about Wang's future. This lost confidence has made it difficult to argue successfully for a system upgrade to solve performance problems or handle user demands. Besides. the cost of such upgrades has grown out of proportion to the benefit.

Other hardware vendors also appear to be in trouble. mergers and strategic partnerships are the order of the day. Industry standards have become fashionable, and proprietary, closed systems, unfashionable. Still, the requirements of your business mandate improved performance. Can important new concepts - client/server architecture, object-oriented design, industry-standard operating systems, non-procedural languages, graphical user interfaces - help?

What to Do?

One viable solution to this dilemma lies in phasing in a new environment to ease the load on the VS while integrating carefully with another system. The new system should allow re-use of present applications while addressing some of the new requirements for integration and data access. Here are a few possible options for this new system:

• PC local area networks: No one would accuse LAN fanatics of restraint in their claims for this mode as a replacement for every other computer option. Nevertheless. it is time to admit that LANs have become a productive solution for many shops using minicomputers. PC-based products and the distribution of data and processing resources are most of the appeal of this approach. LANs offer a good fit for WP shops with little need for large, centralized data storage.

• Other proprietary systems: DEC, IBM, and others have also been noisy in asserting their environments as natural successors to the VS. Like the VS, they are proprietary in structure and inherently host-based in philosophy. Since management may feel burned by their last experience with a proprietary system vendor, they may veto consideration of some valid choices among these vendors.

• UNIX systems: The chameleon of modern systems, UNIX-based machines seem to be large and small, client-based and host-based, modern and primitive. Cheap, powerful and fashionable, they have much of the appeal that VS systems did in the late 1970s and early 1980s. Though suitable for host-based applications, they are more appropriate for networked and distributed processing approaches. UNIX systems are supported by a range of hardware vendors, reducing management anxiety about proprietary systems.

LANs have successfully replaced minicomputers in shops where word processing and office automation are the primary choices, but have not generated as much interest in systems using a common shared database.

Some VS shops should consider the VAX or AS/400 environment simply because of the large amount of applications available. Still, the emerging choice for VS migration seems to be toward UNIX.

UNIX Bound?

Comparisons between the VS and UNIX systems show so few similarities that comparing them is almost comical. The VS offers a common menu-driven environment that requires little training to master. It provides an excellent development environment for business applications. System performance is usually adequate, and there is very little downtime in most shops. VS price/performance remains reasonable by many standards though the cost is much higher today than in the early years. VS applications are usually host-based and distributed applications are rare. As a proprietary operating system, VS/OS is monolithic; that is, most routine operations can be performed from a single viewpoint (the Command Processor) rather than by executing individual functions as needed.

The UNIX user typically faces a command prompt - in fact, it can be one of several types of command prompts. UNIX systems require careful attention by trained personnel to administer. Though commercial applications are catching up, science and engineering software were the first to exploit the UNIX environment. UNIX systems perform well but have a poor reputation for reliability. Small, powerful UNIX systems are cheap and available from a variety of sources. UNIX works best in a networked environment, either as a server or through distribution of processing resources between processors. Finally, effective use and management of UNIX requires an understanding of a large number of commands: what I call a pluralistic approach.

In spite of these glaring differences, some similarities remain:

• UNIX and VS technical users have an independent attitude, preferring to set up and run their systems as they choose, not as some vendor wants them to.

• Most VS users enjoy working with small utilities, such as the USERAIDS collection. UNIX is an operating system composed almost entirely of such tools.

• PC-based software can now act as the front end for UNIX applications, sheltering end users from the operating environment and surpassing the VS's ease of use. UNIX, like the VS, gives users a large amount of control over their environment. This control extends to workstation usage preferences, specialized commands, and management of printers and other peripherals.

In short, it appears for VS sites that initial contact with UNIX will be painful at first, but survival is possible.

The business and personal risk of a change in information management is real and should not be ignored; the benefits are also real and exciting. Your move.

Figure 1: Checklist - Planning for Integration

1. Analyze Needs

   1. System performance
   2. Data distribution to users
   3. Integration with PCs

2. Evaluate Options

   1. System performance
   2. End-user terminal choices (PCs, other workstations)
   3. Support services (staff/user training, maintenance)
   4. Costs (training, staff additions, redevelopment, support, financial impact)

3. Choose Implementation Approach

   1. Total redevelopment (language choice; database choice; application topology - client/server, host, LAN; integration with VS)
   2. Purchase specifications (business requirements, hardware requirements, data exchange with VS)
   3. Port software (choose conversion vendor, revise code, plan exchange with VS)
   4. Combination of above methods